Over one-quarter of the adult population in the United States suffers from a diagnosable mental disorder in a given year and an astonishing 41% of 12th graders report some lifetime use of illicit drugs. Despite the societal and personal burden that psychiatric illness presents and the substantial investment in psychiatric drug discovery (albeit significantly less in drug abuse) there is a chronic shortfall in innovative psychiatric drugs. A primary stumbling block in psychiatric drug development is thought to be in the animal models used to screen drugs for treatment efficacy and the target-centric drug discovery approach. The focus on specific mechanistic interventions will prove unsatisfactory if the underlying pathology does not rest in a restricted biological entity but rather in a `system'response to the drug. Data mining techniques are increasingly used to discover predictive in vitro system profiles for a cancer or toxicological responses. Likewise, a system-based orientation to in vivo pharmacology has been suggested as a way to transform psychiatric drug discovery. The purpose of this application is to fully develop a novel in vivo data mining strategy for psychiatric drug research and development we have termed Pattern Array (PA). The behavioral context for PA is exploratory behavior. Extensive ethological, pharmacological and behavior genetics studies in our lab and others have shown that exploratory behavior is i) highly heritable, likely reflecting `hard-wired'brain systems, ii) amenable to mathematical description and high-throughput and iii) information-rich, generating ~105 relevant data points per animal. Our working hypothesis is that the effects of drugs on this `hard-wired'system are also amenable to algorithmic structuring and identification. The strategy we are proposing is certainly unconventional, however its foundation is well-grounded empirically and shown to work in preliminary studies. We propose to establish PA via three specific aims. First, we will develop a high quality database derived from five main therapeutic target areas: antipsychotics, antidepressants, anxiolytics, drugs of abuse and drug abuse therapeutics. Within each target area a range of subclasses and mechanisms are represented. Second, the strong core database and experience with new drug classes will provide the critical mass to enable us to boost the power, generality and reliability of PA through feature enhancements and statistical implementation. Third, we will utilize PA to mine potential behavioral "endpoints" (~100,000) and identifying those that best characterize a drug or drug class. These endpoints represent complex movement patterns, algorithmically defined as different combinations of several ethologically-relevant variables. The result of this last specific aim will be to provide a set of in vivo behavioral `predictors'for a broad range of compounds with psychoactive properties and provide a template for use in screening novel compounds. PA could then be used to screen novel pharmacotherapeutics for their similarity to proven therapeutics, thus providing a relatively rapid means to identify new molecular entities with unique therapeutic utility. PUBLIC HEALTH RELEVANCE: There is a significant decline in psychiatric drug development designed to treat the considerable portion of the US population that suffers from a diagnosable mental disorder or drug abuse. The purpose of this application is to fully develop an unconventional, novel in vivo data mining strategy for the behavioral effects of psychotherapeutic drugs termed Pattern Array (PA). The strategy outlined in this application could be used to screen novel compounds for their similarity to proven therapeutics, thus providing a relatively rapid means to identify new molecular entities with unique therapeutic utility.